Erratic temp gauge
 

Hi Folks,

For a while now the temp gauge in my 82 240D auto has been screwy.

I believe the gauge is giving false high readings. Everything seems to operate as normal until I have driven for roughly 20-30 minutes. Then the gauge begins to creep up to the 100 degree mark. If i tap on the gauge cluster the needle will drop down to the normal operating temp range right above 80 degrees.

Today on my drive after some long hill climbs I noticed the gauge was reading around 100, if I tapped on it it only moved slightly making me unsure if the motor was actually running too hot or not.

I checked the gauge against my inferred thermometer to test it accuracy and the hottest point on the engine was 78 degrees, lower coolant hose wasn't even hot, leading me to believe the motor wasn't warm enough to warrant the thermostat opening.

Is it possible that higher engine speed means a more intense electrical signal is being sent to the gauge cluster making the gauge go haywire?

I believe the engine is operating within the correct temp range but would like to be able to rely on the temp gauge.

The temp gauge on the W123 cars are known to go haywire after so many years of use. Changing the gauge (temp, oil,fuel) is a pretty simple job. A good used gauge costs about $40-50.

Something else to look at; potential ground issues with the instrument cluster. I had issues with my gauges sporadically flicking around, including one time when the whole cluster went dead for a moment, then came back a few seconds later. I performed the repair mentioned below, and viola, no more gauge flickering, save for a bouncing speedo around the 45-55 mph marks, but as I understand it, swapping for a new speedo cable will fix that.

Mercedes 300D 240D w123 Gauge Cluster Short DIY Fix | Up-Fixin | The Motoring Journal

It's kinda hard to see what's going on in this Internet article's picture, but once you get the cluster out, you can clearly see where to attach the new ground wire, both to the cluster and the ground stud (bolt). You'll see what seems to be about 100,000 brown ground wires coming to one location (the ground bolt). That's the chassis ground point; the other can be seen in the pic above.

Another common cause of erratic gauge needles is caused by a bad ground connection in the gauge itself. They use a long brass rivet to hold it together and to provide a ground. Over time the brass oxidizes and develops a bad connection which causes the needle to bounce.
A fix is to bypass the rivet with a wire going to the gauge shell from the connection on the small circuit board (that is part of the gauge).

A fine tipped soldering pencil (and steady hand) is required for the one connection. The other to the gauge shell will require quite a bit more heat. Also some shells have a gold chromate coating that might need to be filed or sanded off to get a good solder connection. If doing this mod, go ahead and do the fuel gauge as well as it suffers from the same defect.

While you have the cluster apart, resolder the ground connection shown in the 3rd photo. It provides ground to the right side of the cluster and develops cracks as shown. Best to lay a fine piece of wire across to give it extra strength.

I'll re-reiterate...... Almost all weird electrical issues are ground related. I sure would do the ground fixes as detailed above and see if it gets you where you need to go.

Dan

Ditto.....Check the battery-chassis and the chassis-engine grounds. Also, check the connection on the temp. sensor itself before taking things apart or modification.

I had to cut damaged wire off the plug,and resolder new wire,I spliced.going to temp guage and glow plug temp sensor on my 617. the plug comes apart carefully.

This is a pretty good indication that the problem is in the instrument cluster.

Those brass rivets not only oxidize but also become loose with time as they are clamping plastic bits which may deform as they age.

I'll have to check all the above out this weekend and report back

Did you get lost? How’d It work out?!
I did all the things in this post about a year ago and still have issues.
After running new wires to the gauge shells, the bouncing stopped but my fuel gauge is generally off by 15% and the temp sensor rises 10 degrees too hot unless I knuckle the cluster

At the risk of offending Rick, I'm going to borrow one of his photos. The coil that I've circled damps the movement of the meter. If the plastic core is cracked or otherwise compromised, the coil can ground to the metal gauge case. This will produce erratic readings. The quick & dirty is to use electrical tape or heat shrink to insulate the coil. The permanent solution would be to replace the gauge. Just one more thing to check.

Now we're talking.

As it should happen, both of those little plastic spools are cracked at the top (or bottom of the spool). So to be clear, am I placing a stripe of electrical tape on the shell so that they can't bounce into it?

That would do it as a temporary fix.

Gauge Weirdness
 

Thanx Frank ! that's the sort of good tech tip this board thrives on....

I too am always looking for poor grounds and/or dirty / corroded connections before getting to worried and throwing parts at the problems .

" K.I.S.S. ! " . :rolleyes:

Don't mistake the cracks from the actual factory slits on the spool to allow the wire to go through. But if your spool is loose, then just get some super glue and glue it down so it won't flop into anything else.


.

This..... Is the source of all (or most) ground issues in a w123 instrument cluster....



.

Poor Grounds
 

That metal tap is grotty too ~ I always polish connectors that are not shiny .

DO NOT abrade them with sand paper etc. ! . use metal polish and clean rags .

Yep, I reflowed that some time ago and definitely helped. The classic ground wire to the chassis did nothing to stop my bouncing gauges. It was soldering the wire to the shell that finally did the trick... unfortunately it also introduced some new, erratic behavior. Will give reattaching the spool a look in the morning and report back.

Are these gauges and issues common to other models of 81-85?

No real effect... the previous soldering job from rivet to temp gauge shell had come loose. I soldered it again but doesn’t seem to be doing much. For what it’s worth, the issue only occurs around operating temp. If I slap the cluster while she’s warming up, nothing happens.

Think I may have reached the end.

Who’s got a spare 14 pin fuel/temp/oil gauge they’d like to sell me?

Okay, I went beast mode and solved this.

Pulled the cluster once more. I’ve long suspected whatever alloy the temp gauge shell is rolled from is unfriendly to solder. I wrapped a thick copper wire (speaker wire) around the shell and twisted it tight till I couldn’t move it with pliers. Soldered a wire from this ring to the spot on the board. It’s hideous. Shameful in fact.

That said, it’s 82 degrees and cool.

See above.

This was likely true of the fuel gauge -The golden shell sanded down to a dull silver. The temp gauge however was silver already, and while it may have had a coating, I gave it a fairly good sanding to no avail. In any event, i’m sure you’re right. There must’ve been something preventing a good connection.

Maybe not enough heat. I use a Weller 100 watt gun for something like that.

That sounds about right.
My iron (my first to be fair, bought for fixing the clock two years ago) was $5 off a dusty shelf in Napa. It's about the heat of a 100 watt lightbulb. Perhaps time to upgrade.

Just wanated to add my anecdote.
My 1979 300D had a bout of bouncing temp gauge when it reached 150F.

A parts schematic I had from "Performance parts" shows a "thermal vacuum valve" for certain models located on the passenger side of engine block.

Curiously, this bouncing gauge of mine prefaced my door locks beginning to fail.

Then my engine would not shut off periodically.

I had to block off the door lock vacuum lines at the firewall and coincidentally, the temp gauge became steady over a period of 3 days.


I had my mechanic, David Pham, repair the entire vacuum system.
Door locks work, engine shuts off and yes, the temp gauge works fine.

No bouncing.

So I did not bother to check the ground solder point.

(Yes that point does crack, as my spare temp gauge displays the fine crack as the previous photo shows.)

So, perhaps a vac problem can contribute to the temp gauge funkiness via the thermo vacuum valve inefficiency. (?)

Your 'Benz is clearly haunted :rolleyes: .

http://www.peachparts.com/shopforum/members/rick76.html

I would like to extend a sincere thank you for your pictures and explanation of where to find the broken solder joint. I was able to get it diagnosed very quickly and back to normal operation after reflowing the joint and another ground trace.

How timely indeed! My gauge just started the erratic +10-15'C high readings last week. I've been wanting to pull the cluster to replace some bulbs and do the odometer fix for a while now. Good to know there is a good 'while I'm in there' quick fix that can be done!

BE CAREFUL !
 

I occasionally get totally lost in the "Might As Wells" when tinkering with one of my jalopies .

I'm struggling with a temperature gauge problem as well on my 1984 300SD. Mine reads 10C or so low when warmed up. I've replaced the thermostat, the temperature sender, verified correct engine temperature next to the gauge, verified the sender resistance is correct at temperature (59 ohms, which matches what I'd expect at 85C engine temp), but the gauge reads 70C. Today I measured the wire resistance from the sender to the connector on the back of the instrument cluster (pin 2, green wire), and behold, the wire resistance is 17.8 ohms. I resoldered the pin on the engine end, and verified that even if I use a needle to penetrate the insulation on either end of the green wire going from sender to cluster, I get the same reading. It's about the right resistance to explain the low temperature reading since this would make the effective resistance 78 ohms, which is enough to make the gauge read low. I've already added another ground wire to the cluster, and checked the ground resistance from the gauge casing, which all read in the 0.2 ohm range
Any ideas other than running a new wire?

The case is solv-ed. The resistance on the temperature sender wire was caused by a loose connection next to the round interface port on the LF fender (OBD-1?). The wire goes from the temperature sender, to the connector, then up through the fuse box, then to the cluster. Unplugging the connector and reseating eliminated the 17.8 ohm resistance. Temp gauge works perfectly.

Solved & Followup
 

THANK YOU for letting us know the problem found and cure taken .